The servo-type strip-chart recorder error of nonlinearity is specified as less than 0.5 

 percent of full scale. Worst case zero shift and sensitivity change due to temperature varia- 

 tion of 20-degrees Fahrenheit is 2.5 percent of full scale. Worst case zero shift and sensi- 

 tivity change due to line voltage variation is 1.5 percent of full scale for a 10-percent change 

 in line voltage of 1 15 volts, 60 Hertz. However, both the temperature and voltage variation 

 effects will be shown in the following sections to have no effect on final data accuracy. The 

 readout resolution of the recorder is approximately 0.5 percent of full scale based on a speci- 

 fied ink-trace width of 0.25 millimeter which is 0.5 percent of full scale. During basin tests 

 of the BIAS buoy, digital readout devices with readout resolutions of 0.1 percent of full 

 scale were used in conjunction with the recorder to eliminate the human error of bias and 

 interpretation of the recorded output. However, due to limited space on the submarine only 

 the recorder was allowed as a readout device. 



Except for the sensor accuracies, the major contributing factor to overall system error 

 appears to be the nonlinearity of the telemetry electronics and recorder which, when summed 

 (a worst case condition), equal 0.8 percent of full scale. However, a nonlinearity error of 

 this magnitude has not been experienced in any of the calibrations for this system and, further- 

 more, this is a systematic error for which the final data could be corrected to within the error 

 band of readout resolution. 



A remotely-controlled electrical caUbration circuit incorporated in the buoy electronics 

 allows for the direct wiring of the sensor to the resistive calibration elements as depicted in 

 Figure 8. As will be shown, when a voltage is applied to this circuit there exists an invariant 

 voltage ratio between the calibration network outputs and the sensor output which is unaffected 

 by external signal-amplifying electronics. 



For a clear understanding of the principles involved in this technique the discussion will 

 be limited to the buoy pitch measurement channel which contains a potentiometric-type 

 viscous-damped-pendulum sensor. However, these same principles apply to all the remaining 

 measurement channels in this system. The physical calibration and data reduction procedures 

 are explained and an evaluation of the measurement channel is presented. 



CALIBRATION PROCEDURE 



The object of calibrating the pitch sensor is to determine the linearity, sensitivity, and 

 hysteresis, and to determine the angle equivalents of the electrical calibration steps in order 

 to establish the ratio between these steps and the sensor output. Physical calibration requires 

 the use of a tilt table with a precision angle scale so the sensor may be rotated through known 

 positive and negative angles where the horizontal plane is defined as zero degrees. The pendu- 

 lum is affixed to the tilt table and electrically connected to the calibration network, a power 

 supply, and a readout device such as a digital voltmeter (DVM) with a readout resolution and 



22 



